Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Effects of pulsatile blood flow in large vessels on thermal dose distribution during thermal therapy Tzyy-Leng Horng Department of Applied Mathematics, Feng Chia University, Taichung, Taiwan Win-Li Lin Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan and Medical Engineering Research Division, National Health Research Institutes, Miaoli, Taiwan Chihng-Tsung Liauh Department of Mechanical Engineering, Kun Shan University of Technology, Tainan, Taiwan Tzu-Ching Shiha兲 Department of Medical Radiology Technology, China Medical University and Department of Radiology, China Medical University Hospital, Taichung, Taiwan 共Received 1 November 2006; revised 3 February 2007; accepted for publication 6 February 2007; published 16 March 2007兲 The aim of this study is to evaluate the effect of pulsatile blood flow in thermally significant blood vessels on the thermal lesion region during thermal therapy of tumor. A sinusoidally pulsatile velocity profile for blood flow was employed to simulate the cyclic effect of the heart beat on the blood flow. The evolution of temperature field was governed by the energy transport equation for blood flow together with Pennes’ bioheat equation for perfused tissue encircling the blood vessel. The governing equations were numerically solved by a novel multi-block Chebyshev pseudospectral method and the accumulated thermal dose in tissue was computed. Numerical results show that pulsatile velocity profile, with various combinations of pulsatile amplitude and frequency, has little difference in effect on the thermal lesion region of tissue compared with uniform or parabolic velocity profile. However, some minor differences on the thermal lesion region of blood vessel is observed for middle-sized blood vessel. This consequence suggests that, in this kind of problem, we may as well do the simulation simply by a steady uniform velocity profile for blood flow. © 2007 American Association of Physicists in Medicine. 关DOI: 10.1118/1.2712415兴 Key words: pulsatile blood flow, bioheat transfer, thermal dose, Chebyshev pseudospectral method I. INTRODUCTION Blood flow can remarkably affect the temperature distributions in tumor tissues during thermal therapies, especially with the presence of large blood vessels.1–6 Thermally significant blood vessels 共larger than 0.2 mm in diameter兲 cool tissue during thermal treatments, making it very difficult to cover the whole tumor volume with therapeutic thermal exposure.7–9 Shih et al.9 found that the short-duration highintensity heating can treat a target tumor inside with a blood vessel diameter of 0.2 mm when using the total power energy density up to 120 J cm−3. Moreover, Becker and Kuznetsov7 found that the target region tissue was found to have a lower thermal damage in the presence of a blood vessel, and this cooling effect is more pronounced as the blood vessel gets larger 共larger than 0.6 mm in diameter兲. These phenomena motivated us to thoroughly study the effects of thermally significant blood vessels on thermal therapy of tumor. The cyclic nature of the heart pump generates pulsatile blood flow in all arteries.10–20 This periodic-in-time velocity profile for pulsatile blood flow, driven by an oscillating pressure gradient, inside a circular blood vessel was first studied by Womersley.20 Loudon and Tordesillas21 used dimensionless Womersley number to characterize the pulsatile fre1312 Med. Phys. 34 „4…, April 2007 quency of blood flow. They showed that when Womersley number is small, the flow tracks the oscillating pressure gradient faithfully and the velocity profile exhibits a parabolic shape. Nevertheless, when Womersley number is large, the phase difference between velocity and pressure gradient becomes larger and the velocity profile exhibits a shape of two peaks. Generally in hyperthermia treatment of tumors, blood vessels with a diameter larger than 0.5 mm can no longer be included in blood perfusion and must be treated individually22–26 because temperature difference between blood and surrounding tissue is significant. Heat transfer in tissue within the region near a large blood vessel is especially influenced significantly by the flowing blood.27,28 Furthermore, Mohammed and Verhey28 utilized a finite element method to simulate laser interstitial thermo therapy on anatomical inhomogeneous regions by considering blood perfusion only and also performed in in vitro experiment on porcine muscle tissue. Their numerical results had 5% to 20% discrepancy in thermal lesion compared with the experiment. This indicates the importance of an appropriate thermal model to achieve reliable results from numerical simulations. So far no numerical result considering the coupled heat transfer problem of both a large blood vessel with pulsatile 0094-2405/2007/34„4…/1312/9/$23.00 © 2007 Am. Assoc. Phys. Med. 1312 1313 Horng et al.: Pulsatile blood flow blood flow inside and its surrounding tissue is ever reported, and the whole physics remains poorly understood. Craciunescu and Clegg15 numerically solved the axis-symmetric Navier-Stokes equations to obtain the pulsatile velocity profile and then computed the energy transport equation to obtain the temperature distribution of pulsating blood flow in a blood vessel with sizes of aorta, large arteries, terminal arterial branches, and arterioles. Some cases of their simulation have pulsation of large amplitude such that reversal of blood flow occurs in large vessels 共i.e., 3 mm in diameter and larger兲. However, they only focused on the temperature distribution within a blood vessel and did not consider that for its surrounding tissue. Shih et al.8 computed the coupled heat transfer problem of both a large blood vessel and its surrounding tissue. However, they only assumed the velocity profile of the blood flow inside the vessel to be simply steady and uniform. They found that the thermal dose would be insufficient for treatment in the region near blood vessel when the radius and flow rate are large. Shih et al.9 extended Shih et al.8 to investigate the influence of various heating speeds with the total heat applied held equal and found the thermal dose deficit in the region near blood vessel is less severe when the heating speed is large. Here we studied this coupled heat transfer problem between a thermally significant blood vessel and its surrounding tissue during heat treatment by considering the velocity profile of the blood flow being uniform, parabolic, and pulsatile under various radii of vessel, pulsating frequencies and amplitudes 共in the case of pulsatile velocity profile兲, and heating speeds. The main object is to realize the effects of flow velocity profile, size of vessel, flow arte, and heating speed on thermal dose distribution of the surrounding tissue during heat treatment. Though many factors about blood vessels have been brought into consideration in the current study compared with existing literatures, it is still far from being complete. Many other factors such as vessel wall elasticity, porosity, rheology of blood, realistic shape of blood pressure wave, possible counter current flow, and even turbulence in pulsatile blood flow are either ignored or simplified in the current study.11,12,29 Here the whole paper is organized as follows: the mathematical model and its governing equations followed by its numerical method are described in the section of mathematical model and numerical method. The numerical results are presented in the section of results and discussions with the effects of flow velocity profile, size of vessel, flow rate, and heating speed on thermal dose distribution discussed. The final section is conclusion to summarize all key physics in this study. II. MATHEMATICAL MODEL AND NUMERICAL METHOD A. Velocity profile of pulsatile blood flow in a circular blood vessel Besides being simply steady uniform or parabolic, the pulsation of blood flow is considered here and its velocity profile is derived as follows. Here, we assume the blood flow to be incompressible, laminar and Newtonian obeying Medical Physics, Vol. 34, No. 4, April 2007 1313 FIG. 1. Velocity profile of pulsatile blood flow at different phases for r0 = 0.5 mm, f = 1 Hz, ␣ = 0.64213, fac= 0.2. 共a兲 oscillating part of total velocity, 共b兲 total velocity. Navier-Stokes equations in a cylindrical impermeable blood vessel with rigid wall of radius r0. The axial HagenPoiseuille steady parabolic velocity profile is12 w=− 1 2 2 dp 共r − r 兲 . 4 0 dz 共1兲 For pulsatile flows, the pressure gradient above is no more a constant, and is in fact periodic in time due to the rhythmic nature of beating heart. To mimic the realistic shape of pressure gradient in time, it requests at least 11 harmonics,29 and here we simply consider the fundamental one only for the feasibility of analysis. With this additional sinusoidal component in time, the corresponding pressure gradient, without loss of major physics, along the z-axis of blood vessel can be assumed periodically as p = c0 + c1 sin t z = c0 + c1 eit − e−it 2i = c0 + ĉ1eit + ĉ−1e−it , 共2兲 where = 2 f is the angular frequency; the period is denoted as T = 1 / f; ĉ1 = c1 / 2i, ĉ−1 = −c1 / 2i. The corresponding axial 1314 Horng et al.: Pulsatile blood flow 1314 velocity profile W共r , t兲 can be similarly expressed in the form as W = w + w1eit + w−1e−it , w̄ = 共3兲 冉 冊 by Eq. 共1兲. The volume flow rate with averaged over the period T is Q̇avg = 2 T 冕冕 0 r0 Wrdrdt = − 0 r40 8 W = 2w̄ 1 − 共4兲 c0 , 冉 冊 r2 + w1eit + w−1e−it r20 冤 冤 冉 冉 冊 冊 冥 冉 冊 冤 冉 冊冥 冉 冊 冦冤 冉 冊冥 冧 冉 冊 r J0 ␣ i3/2 r0 iĉ1 + 1− r J0 ␣ i3/2 r0 冉 冊 r J0 ␣ i3/2 r0 c1 + 1− r 2 J0 ␣ i3/2 r0 2 r = 2w̄ 1 − 2 r0 2 r = 2w̄ 1 − 2 r0 冉 冊 r2 c1 = 2w̄ 1 − 2 + Re r0 冤 r0 denotes the Womersley number; and the density and kinematic viscosity of blood, respectively; Re共·兲 the real part; J0 is the Bessel function of the first kind of order zero. Also, we denote here 冒冉 冊 − 8w̄ r20 to characterize the relative intensity of pulsation in blood flow. In current study, the value of fac ranges from 0.2 to 1, that is not too large to cause reverse flow judging from possible practical situations. The Womersley number ␣ is used to describe the competition between transient inertia force and viscous force. If the Womersley number ␣ is large, the effect of viscosity cannot propagate far from the wall, and the flow in the central portion of tube acts like inviscid flow and is chiefly determined by the balance of inertia force and pressure gradient. On the other hand, when the Womersley number ␣ is large, the velocity profile in a pulsatile flow is Medical Physics, Vol. 34, No. 4, April 2007 冥 共6兲 e−it 冊 冊 冥 e−it e it , 冑/ fac = c1/c0 = c1 冉 冉 冊 r J0 ␣ i3/2 r0 c1 e it + 1− r 2 J0 ␣ i3/2 r0 where ␣= 冉 r J0 ␣ i5/2 r0 iĉ−1 e it − 1− J0共␣i5/2兲 r J0 ␣ i3/2 r0 1− r J0 ␣ i3/2 r0 r2 + w1eit + w−1e−it . r20 The sinusoidal components w1 and w−1 can be determined by solving Navier-Stokes equations, and Eq. 共6兲 can be expressed as follows by Fung:12 and the average velocity in space and time is, therefore, W = 2w̄ 1 − 共5兲 The velocity profile can then be further expressed as w = −1 / 4共r20 − r2兲c0 T c0r20 Q̇avg = − . 8 r20 共7兲 expected to be relatively blunt compared with the steady parabolic profile of Poiseuille flow. When Womersley number ␣ is larger, the pulsating velocity profile may even have two peaks. However, this two-peak velocity profile did not occur in the current study considering the vessel diameter used here ranging from 0.2 to 2 mm and heart beat frequency 1 to 2 Hz that makes maximum Womersley number 1.8162 only. Figures 1共a兲 and 1共b兲 show the velocity profile of oscillating component and the total, respectively, for various phases in a complete cycle for a typical case in the current study 共r0 = 0.5 mm, f = 1 Hz, ␣ = 0.642 13, fac= 0.2兲. Neither two-peak behavior nor reverse flow is observed in the velocity profile. B. Governing equations Here we assume that the absorbed power density in blood and tissue is equal to the heating power density. The axissymmetric geometric configuration considered here is a cylindrical perfused tissue, including tumor and normal tissues, with a coaxial rigid blood vessel inside and throughout the tissue as shown in Fig. 2. The whole computational domain is bounded by r = rmax, z = 0, and z = zmax; the blood vessel is 1315 Horng et al.: Pulsatile blood flow 1315 TABLE II. Three different heating schemes used in current simulations. Case Absorbed power density Q 共W cm−3兲 Heating duration th 共s兲 Total absorbed energy density 共J cm−3兲 FIG. 2. Geometric configuration in current simulations. 共a兲 The treatment target 共heating target兲 is specified as z1 艋 z 艋 z2, 0 艋 r 艋 r1, with z1 = 5 mm, z2 = 15 mm, r1 = 5 mm considered here. 共b兲 Schematic illustration of the three kinds of velocity profile of blood flow in blood vessels. Left-steady uniform velocity profile, middle-steady parabolic velocity profile, rightpulsatile velocity profile. surrounded by r = r0, z = 0, and z = zmax; the heating target 共tumor and a part of blood vessel inside the tumor兲 is bounded by r = r1, z = z1, and z = z2. The diameters of blood vessels and their associated average flow velocities considered in the current study are presented in Table I. The governing equations for the temperature evolution are energy transport equations shown in Eqs. 共8兲 for tissue and 共9兲 for blood vessel: tc t I II III 50 2 100 10 10 100 2 50 100 and specific heat, respectively, that are all assumed to be constant; Q共r , z , t兲 is the power of heat added axissymmetrically; Wb is the perfusion mass flow rate; Ta is the ambient temperature that is usually set to be 37 ° C; w共r , t兲 is that axial velocity of blood flow; subscripts t and b represent tissue and blood, respectively. Equation 共8兲 is in fact Pennes’ bioheat transfer equation30 in cylindrical coordinates for tissue with the heat sink −Wbcb共Tt − Ta兲 to describe the perfusion effect by the network of microvascular blood flow. Equation 共9兲 is the conventional energy transport equation for blood vessel with both convection and diffusion taken into account. The initial condition is Tt共r,z,0兲 = Tb共r,z,0兲 = Ta = 37 ° C. 共10兲 At the interface conditions between the blood vessel and tissue are continuity of temperature and heat flux Tt = Tb kt at ⌫, Tt Tb = kb n n 共11兲 at ⌫, 共12兲 冋 冉 冊 册 1 Tt Tt 2T t = kt r + 2 − Wbcb共Tt − Ta兲 t r r r z + Qt共r,z,t兲, bc b 冉 冊 冋 冉 冊 册 共8兲 1 Tb Tb Tb 2T b +w = kb r + 2 + Qb共r,z,t兲, z t r r r z 共9兲 where T共r , z , t兲 denotes the temperature that is distributed axis-symmetrically; , k , c are density, thermal conductivity, TABLE I. List of the blood vessel parameters used in current simulations 共Ref. 8兲. Diameter 共mm兲 Average blood velocity in tumor 共w̄兲 共mm s−1兲 0.2 0.6 1.0 1.4 2.0 3.4 6 8 10.5 20 Medical Physics, Vol. 34, No. 4, April 2007 FIG. 3. The computational domain is decomposed into nine rectangular blocks. Blocks 1, 4, and 7 are blood vessel; block 5 is the tumor tissue; the other blocks are normal tissue. Blocks 4 and 5 are heating zone. Here the mesh resolution is 10⫻ 10 in each block. Due to the feature of Chebyshev collocation points, grids are clustered more densely near boundary/interface in each block. In this figure rmax = 10 mm, zmax = 100 mm, r0 = 2 mm, r1 = 5 mm, z1 = 10 mm, z2 = 20 mm. 1316 Horng et al.: Pulsatile blood flow 1316 FIG. 5. Effect of the frequency of the pulsatile blood flow on the thermal lesion region 共shaded region兲 for the blood vessels 共a兲, 共f兲, 共k兲 0.2 mm; 共b兲, 共g兲, 共l兲 0.6 mm; 共c兲, 共h兲, 共m兲 1 mm; 共d兲, 共i兲, 共n兲 1.4 mm; 共e兲, 共j兲, 共o兲 2 mm in diameter. 共a兲–共e兲 with the frequency 1 Hz, 共f兲–共j兲 with the frequency 1.5 Hz, and 共k兲–共o兲 with the frequency 2 Hz. The blood vessel boundaries are denoted with the horizontal dashed lines. The heated target region is denoted by a square with dashed lines. Here r1 = 5 mm, Wb = 2 kg m−3 s−1, b = t = 1050 kg m−3, cb = ct = 3770 J kg−1 ° C−1, kb = kt = 0.5 W m−3 ° C−1. The target is heated in the way of the heating case I 共i.e., Qt = Qb = 50 W cm−3, and the heating duration= 2 s兲 in Table II. FIG. 4. Effect of the steady velocity profiles of blood flow on the thermal lesion region 共shaded region兲 for the blood vessels: 共a兲, 共f兲 0.2 mm; 共b兲, 共g兲 0.6 mm; 共c兲, 共h兲 1 mm; 共d兲, 共i兲 1.4 mm; 共e兲, 共j兲 2 mm in diameter. 共a兲–共e兲 are results of a uniform velocity profile, and 共f兲–共j兲 are results of a parabolic one. The blood vessel boundaries are denoted with the horizontal dashed lines. The heated target region 共tumor兲 is denoted by a square with dashed lines. Here r1 = 5 mm, Wb = 2 kg m−3 s−1, b = t = 1050 kg m−3, cb = ct = 3770 J kg−1 ° C−1, kb = kt = 0.5 W m−3 ° C−1. The target is heated in the way of the heating case I 共i.e., Qt = Qb = 50 W cm−3, and the heating duration = 2 s兲 in Table II. where ⌫ denotes the interface between blood vessel and tissue, and n denotes the direction normal to ⌫. At r = 0, pole condition is applied for blood vessel by axis-symmetry, Medical Physics, Vol. 34, No. 4, April 2007 Tb = 0. r 共13兲 Finally, the boundary condition at r = rmax, z = 0, and z = zmax is temperature all equal to the ambient one, Tt 共or Tb兲 = Ta = 37 ° C, 共14兲 except the blood part at z = zmax using convective boundary condition, Tb Tb = 0, +w z t at z = zmax . 共15兲 1317 Horng et al.: Pulsatile blood flow 1317 FIG. 6. Comparison of the effect of various velocity profiles on the thermal lesion region for 1-mm-diam blood vessel with 共a兲 steady uniform velocity profile, 共b兲 steady parabolic velocity profile, 共c兲 1 Hz and fac = 0.2 pulsatile velocity profile, 共d兲 1.5 Hz and fac= 0.2 pulsatile velocity profile, 共e兲 2 Hz and fac= 0.2 pulsatile velocity profile. C. Calculation of thermal dose The accumulated thermal dose to tissue is a function of heating duration and the temperature level. The estimate of tissue damage is based on the thermal dose of which the formula was proposed by Sapareto and Dewey.31 The thermal dose or equivalent minutes at 43 ° C 共EM43兲 is shown as follows: EM43共in min . 兲 = 冕 tf RT−43dt, 共16兲 0 where R = 2 for T 艌 43 ° C, R = 4 for 37 ° C ⬍ T ⬍ 43 ° C, and t f = 90 s or 120 s in current simulation. The threshold dose for necrosis is EM43 = 240 min for muscle tissue.32 Therefore, the region encircled by the level curve EM43 = 240 min is taken as the thermal lesion region in this study. Covering tumor tissue but not normal tissue by thermal lesion region as full as possible is most desired in the thermal treatment, Medical Physics, Vol. 34, No. 4, April 2007 and the deficit would serve as an evaluation of the effectiveness of the treatment. This region is greatly influenced by size of blood vessel31 and the heating speed.32 Here three different heating schemes characterizing different heating speeds under the same amount of heat added 共100 J cm−3 from preliminary energy analysis in lump33兲 are depicted in Table II. D. Numerical method Here we used method of lines 共MOL兲 to solve Eqs. 共8兲–共15兲. First, we employed multi-block Chebyshev pseudospectral method to discretize governing equations 共8兲 and 共9兲, boundary and interface conditions 共10兲–共15兲 in space into a semi-discrete system in time. This coupled system consists of ordinary differential equations in time, mainly from Eqs. 共8兲 and 共9兲, and algebraic equations, chiefly from boundary and interface conditions 共10兲–共15兲. We can then 1318 Horng et al.: Pulsatile blood flow 1318 FIG. 9. Effect of fac of the pulsatile blood flow on the thermal lesion region with the pulsatile frequency being 1 Hz. The target is heated in the way of the heating case I 共i.e., Qt = Qb = 50 W cm−3, and the heating duration= 2 s兲 in Table II. 共a兲 fac= 0.2, 共b兲 fac= 0.5, 共c兲 fac= 0.8, and 共d兲 fac= 1. FIG. 7. Comparison of the effect of different heating schemes and blood vessel diameters on the thermal lesion region with steady uniform velocity profile. The solid and dashed lines represent the heating cases I and II, and the shaded area represents the heating case III. 共a兲 The diameter of the blood vessel is 0.2 mm. 共b兲 The diameter of the blood vessel is 1 mm. solve this system of differential-algebraic equations 共DAE’s兲 by many efficient DAE solvers available from most numerical libraries, and here we used popular MATLAB index-1 DAE solver ode15s in our computation. Multi-block Chebyshev pseudospectral method first decomposes the computational domain into several blocks, based on heterogeneity of thermal properties or heating, and discretizes equations in space in each block by conventional Chebyshev pseudospectral method. Details about Chebyshev pseudospectral meth- FIG. 8. The temperature contours for the case of blood vessel diameter 1 mm with steady uniform blood flow under case-I heating 共i.e., Qt = Qb = 50 W cm−3, and the heating duration= 2 s兲 at the time of power off 共t = 2 s兲 with the highest temperature reaching 62.108 ° C. Medical Physics, Vol. 34, No. 4, April 2007 ods are referred to Canuto et al.,34 Trefethen,35 and Fornberg.36 Here the computational domain is decomposed to nine rectangular blocks to cope with blood vessel and heating zone as shown in Fig. 3. We can observe from Fig. 3 that the grids are clustered more densely near boundary/ FIG. 10. Effect of fac of the pulsatile blood flow on the thermal lesion region with the pulsatile frequency being 1.5 Hz. The target is heated in the way of the heating case I 共i.e., Qt = Qb = 50 W cm−3, and the heating duration= 2 s兲 in Table II. 共a兲 fac= 0.2, 共b兲 fac= 0.5, 共c兲 fac= 0.8, and 共d兲 fac= 1. 1319 Horng et al.: Pulsatile blood flow 1319 interface due to the characteristics of Chebyshev collocation points.34–36 This feature makes the computational accuracy and efficiency far better than the one without any decomposition. In short, owing to the intrinsic high order of accuracy of pseudospectral methods, grid size is not that restricted as finite difference/element methods and, hence, the overall computation is very efficient. III. RESULTS AND DISCUSSIONS Figure 4 compares the effects of steady uniform and parabolic velocity profiles for blood flow on thermal lesion region with heating scheme I in Table II, which has the least thermal dose deficit as described in Shih et al.9 Likewise, Fig. 5 compares the effects of pulsatile blood flow with various pulsatile frequencies 共1, 1.5, and 2 Hz兲 with the relative intensity of pulsation fac= 0.2 on thermal lesion region with heating Scheme I in Table II. Figures 4 and 5 generally show that there is almost no difference in thermal lesion region among all these velocity profiles under the same size of blood vessel. Only minor difference of thermal lesion region in blood vessel is observed in middle-sized blood vessels 共with diameter 0.6 and 1 mm兲. This minor difference is most pronounced for blood vessel of diameter 1 mm and is particularly shown in Fig. 6. It shows that the length of thermal lesion region in blood vessel is longest for uniform velocity profile, shortest for parabolic and 1.5 Hz pulsatile ones, and in between for 1 and 2 Hz pulsatile flows. While the thermal lesion region is rather insensitive to the velocity profile of blood flow, it is deeply influenced by the size of blood vessel since the heat convection by the blood flow in a blood vessel usually serves as a stronger heat sink than the blood perfusion in tissue. That means temperature would drop faster in a blood vessel than its surrounding tissue. This may cause deficit of thermal lesion region in blood vessel and tissue nearby, which can be easily observed from Figs. 4 and 5. Generally the deficit of thermal lesion region is less for smaller vessels. In the case of smallest vessel here 共with diameter 0.2 mm兲, thermal lesion region almost covers the whole blood vessel inside the tumor and the deficit is naturally the least. For middle-sized vessels here 共with diameter 0.6 and 1 mm兲, the thermal lesion region in blood vessel becomes smaller and shifted downstream. For large vessels here 共with diameter 1.4 and 2 mm兲, there is a total deficit of thermal lesion region in blood vessel, and this would cause deficit in tumor tissue right near the blood vessel especially at the upstream. Besides diameters of blood vessel, the thermal lesion region is also very sensitive to the heating speed. Figure 7 generally shows larger thermal lesion region for faster heating, and the heating speed affects more pronouncedly when the blood vessel is larger. As shown in Fig. 7共b兲, there exists an obvious shift of thermal lesion region to the downstream of blood vessel for the blood vessel with diameter 1 mm when heating is fast, and this may cause unwanted thermal injury in normal tissue nearby. Figure 8 particularly shows the temperature contours for the case of blood vessel diameter of 1 mm under heating Scheme I in Fig. 7 at the time of Medical Physics, Vol. 34, No. 4, April 2007 FIG. 11. Effect of fac of the pulsatile blood flow on the thermal lesion region with the pulsatile frequency being 2 Hz. The target is heated in the way of the heating case I 共i.e., Qt = Qb = 50 W cm−3, and the heating duration= 2 s兲 in Table II. 共a兲 fac= 0.2, 共b兲 fac= 0.5, 共c兲 fac= 0.8, and 共d兲 fac= 1. power off 共t = 2 s兲. Notice that the highest temperature occurs right at the downstream of blood vessel. This causes the thermal damage of the normal tissue nearby. The effect of pulsation amplitude, in terms of relative intensity fac, of pulsatile blood flow on the thermal lesion region generally has little difference among various fac’s, except minor difference for middle-sized blood vessels. Figures 9–11 depicts this effect for a middle-sized blood vessel of diameter 1 mm with fac= 0.2, 0.5, 0.8, and 1 and frequency being at 1 Hz 共Fig. 9兲, 1.5 Hz 共Fig. 10兲, and 2 Hz 共Fig. 11兲, respectively. When fac increases, the blood flows more in a stick-slip fashion, and this may considerably influence the heat convection when incorporated with pulsation frequency. Figures 9, 10, 11共c兲, and 11共d兲, with fac= 0.8 and 1, respectively, show obviously two-peak behavior in thermal dose contour at the downstream of blood vessel, which is chiefly because of large pulsation amplitude. IV. CONCLUSION The current investigation shows that the effect of velocity profile of blood flow, ranging from uniform, parabolic, to pusatile ones, has almost no difference in thermal lesion region on tumor region and minor difference only on blood vessel when the blood vessel is of middle size. This result suggests that we might just as well use the simplest steady uniform velocity profile to do the simulation without significant difference in thermal lesion region of tumor. In fact, the thermal lesion region is much more sensitive to heating speed and size of blood vessel. Some studies8,9,26,33 show that faster heating would form a much better thermal lesion 1320 Horng et al.: Pulsatile blood flow region, and it works best on small blood vessels with a better covering of both the tumor and blood vessel by the thermal lesion region since the heat convection by the blood flow in the blood vessel is least. For large vessels, it has a total deficit in blood vessel and some deficit in tumor near the upstream of blood vessel. As to middle-sized vessels, a shift of partially deficient thermal lesion region to the downstream of blood vessel may cause unwanted thermal injury to the normal tissue nearby. Furthermore, for middle-sized blood vessels, pulsatile blood flow of large pulsation amplitudes may further cause a two-peak behavior in the thermal dose contour at the downstream of blood vessel. ACKNOWLEDGMENT This research was supported in part by the National Science Council of the Republic of China under Grant No. NSC-94-2213-E-039-004. a兲 Author to whom correspondence should be addressed. Electronic mail: [email protected] 1 J. Crezee and J. J. W. Lagendijk, “Experimental verification of bioheat transfer theories: measurement of temperature profiles around large artificial vessels in perfused tissue,” Phys. Med. Biol. 35, 905–923 共1990兲. 2 W. M. Whelan, D. R. Wyman, and B. C. Wilson, “Investigations of large vessel cooling during interstitial laser heating,” Med. Phys. 22, 105–115 共1995兲. 3 M. C. Kolios, M. D. Sherar, and J. W. Hunt, “Large blood vessel cooling in heated tissues: A numerical study,” Phys. Med. Biol. 40, 477–494 共1995兲. 4 M. C. Kolios, A. E. Worthington, M. D. Sherar, and J. W. Hunt, “Experimental evaluation of two simple thermal models using transient temperature analysis,” Phys. Med. Biol. 43, 3325–3340 共1998兲. 5 C. X. Zhang, S. Zhang, Z. Zhang, and Y. Z. Chen, “Effects of large blood vessel locations during high intensity focused ultrasound therapy for hepatic tumors: a finite element study,” Proceeding of the 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference, Shanghai, China, 1–4 September 2005. 6 M. Nikfarjam, V. Muralidharan, C. Malcontenti-Wilson, W. McLaren, and C. Christophi, “Impact of blood flow occlusion on liver necrosis following thermal ablation,” ANZ J. Surg 76, 84–91 共2006兲. 7 S. M. Becker and A. V. Kuznetsov, “Numerical modeling of in vivo plate electroporation thermal dose assessment,” ASME J. Biomech. Eng. 128, 76–84 共2006兲. 8 T. C. Shih, H. S. Kou, and W. L. Lin, “The impact of thermally significant blood vessels in perfused tumor tissue on thermal distributions during thermal therapies,” Int. Commun. Heat Mass Transfer 30, 975–985 共2003兲. 9 T. C. Shih, H. L. Liu, and A. T. L. Horng, “Cooling effect of thermally significant blood vessels in perfused tumor tissue during thermal therapy,” Int. Commun. Heat Mass Transfer 33, 135–141 共2006兲. 10 L. E. Fencil, K. Doi, K. G. Chua, and K. R. Hoffman, “Measurement of absolute flow rate in vessels using a stereoscopic DSA system,” Phys. Med. Biol. 34, 659–671 共1989兲. 11 W. W. Nichols and M. F. O’Rourke, McDonald’s Blood Flow in Arteries: Theoretic, Experimental and Clinical Principles 共Edward Arnold, London, 1990兲. 12 Y. C. Fung, Biomechanics: Motion, Flow Stress, and Growth 共SpringerVerlag, New York, 1996兲. 13 D. K. De, “Theoretical investigation of estimation of steady and pulsatile Medical Physics, Vol. 34, No. 4, April 2007 1320 blood flow and blood vessel cross section by CW NAR excitation,” Phys. Med. Biol. 35, 197–211 共1990兲. 14 D. N. Ku, “Blood flow in arteries,” Annu. Rev. Fluid Mech. 29, 399–434 共1997兲. 15 O. I. Craciunescu and S. T. Clegg, “Pulsatile blood flow effects on temperature distribution and heat transfer in rigid vessels,” J. Biomech. Eng. 123, 500–505 共2001兲. 16 M. B. Robertson, U. Köhler, P. R. Hoskins, and I. Marshall, “Flow in elliptical vessels calculated for a physiological waveform,” J. Vasc. Res. 38, 73–82 共2001兲. 17 B. Smith, J. G. Chase, R. I. Nokes, G. M. Shaw, and T. David, “Velocity profile method for time varying resistance in minimal cardiovascular system models,” Phys. Med. Biol. 48, 3375–3387 共2003兲. 18 L. Yao, D. Liao, Y. Zeng, X. Xu, and H. Xu, “Radial distributions of temperature pressure and velocities for pulsatile blood flow in an axisymmetrical stiff tube,” Physiol. Meas 25, 1437–1442 共2004兲. 19 Y. J. Zeng, J. H. Zhang, B. Shen, Y. Diao, X. H. Xu, and H. Xu, “Measuring blood flow velocities based on three image processing techniques,” Med. Phys. 32, 1187–1192 共2005兲. 20 J. R. Womersley, “Method for the calculation of velocity, rate of flow and viscous drag in arteries when the pressure gradient is know,” J. Physiol. 共London兲 127, 553–563 共1955兲. 21 C. Loudon and A. Tordesillas, “The use of the dimensionless Womersley number to characterize the unsteady nature of internal flow,” J. Theor. Biol. 191, 63–78 共1998兲. 22 A. Kotte, G. van Leeuwen, J. de Bree, J. van der Koijk, H. Crezee, and J. Lagendjik, “A description of discrete vessel segments in thermal modeling of tissues,” Phys. Med. Biol. 41, 865–884 共1996兲. 23 J. C. Chato, “Heat transfer to blood vessels,” J. Biomech. Eng. 102, 110–118 共1980兲. 24 J. J. W. Lagendijk, “The influence of bloodflow in large vessels on the temperature distribution in hyperthermia,” Phys. Med. Biol. 27, 17–23 共1982兲. 25 J. Crezee and J. J. W. Lagendijk, “Temperature uniformity during hyperthermia: the impact of large vessels,” Phys. Med. Biol. 37, 1321–1337 共1992兲. 26 M. C. Kolios, M. D. Sherar, and J. W. Hunt, “Blood flow cooling and ultrasonic lesion formation,” Med. Phys. 23, 1287–1298 共1996兲. 27 L. Consiglieri, I. dos Santos, and D. Haemmerich, “Theoretical analysis of the heat convection coefficient in large vessels and the significance for thermal ablative therapies,” Phys. Med. Biol. 48, 4125–4134 共2003兲. 28 Y. Mohammed and J. F. Verhey, “A finite element method model to simulate laser interstitial thermo therapy in anatomical inhomogeneous regions,” Biomed. Eng. Online 4, 1–16 共2005兲. 29 Y. C. Fung, Biomechanics: Circulation 共Springer-Verlag, New York, 1996兲. 30 H. H. Pennes, “Analysis of tissue and arterial blood temperature in the resting human forearm,” J. Appl. Physiol. 1, 93–122 共1948兲. 31 S. A. Sapareto and W. C. Dewey, “Thermal dose determination in cancer therapy,” Int. J. Radiat. Oncol., Biol., Phys. 10, 787–800 共1984兲. 32 C. Damianou and K. Hynynen, “Focal spacing and near-field heating during pulsed high temperature ultrasound therapy,” Ultrasound Med. Biol. 19, 777–787 共1993兲. 33 H. S. Kou, T. C. Shih, and W. L. Lin, “Effect of the directional blood flow on thermal dose distribution during thermal therapy: An application of a Green’s function based on the porous model,” Phys. Med. Biol. 48, 1577–1589 共2003兲. 34 C. Canuto, M. Y. Hussaini, A. Quarteroni, and T. A. Zang, Spectral Methods, Fundamentals in Single Domains 共Springer-Verlag, Berlin, Heidelberg, 2006兲. 35 L. N. Trefethen, Spectral Methods in MATLAB 共SIAM, Philadelphia, 2000兲. 36 B. Fornberg, A Practical Guide to Pseudospectral Methods 共Cambridge University Press, Cambridge, 1996兲.